[Kostas_K]

Hello,

I am runnig a simulation to visualize a turbulent jet. I solved it as steady state first with k-e model in order to get an initial value and the I use k-omega for the residuals to converge.
I want to visualize the Reynolds Stresses without using the RSM. I looked around and i found that I have to create a custom field function with Boussinesq approximation for calulating Reynolds stresses. I also found the theory behind this and I know how to write simple custom field functions but I can not find how to write the equations in order to calculate them.
Has anyone done this before in order to help me/ guide me ?

Any help is welcome.
Thank you

[LuckyTran]

I have a cute implementation as a custom field function. However, before v18, custom field functions cannot be used in parallel so this can only be used for post-processing and not on the fly. If you want it to work on the fly, you need to do it using UDF and UDM (only for <v18). v18 fluent allows parallel custom field functions.

Plotting of normal stresses from k-e model

Just load the file directly. I think it works for any of the k-based models since all you need is the k, strain rate, and turbulent viscosity. If I get motivated I'll make it a sticky on my blog. Please also be careful of the density appearing in my script. It depends how you define 'stress' whether or not there should be a density involved.

If you know the equations, it's very easy (though cumbersome) to set up the custom field functions in the GUI since you just click on a bunch of buttons. It's much tougher to do manually in text in one go. I would recommend trying to reproduce just one of the equations so you can convince yourself that you know how to do it.

[Kostas_K]

Thank you very much LuckyTran for your response and for your file

I am solving it as incompressible so the density is a constant. Will there be a problem ?

[LuckyTran]

Constant density isn't a problem, you still have a field variable called density and it will still work.

The density issue is how do you define the Reynolds stresses? If you take the terms out of the momentum equation with units of stress then it should be rho*u'*u'. However, many people define the uu Reynolds stress as u'*u' with no density. The question is how do you define it? You can keep or delete the density to suit your needs.

[Kostas_K]

Well what i exactly want to visualize are the Reynolds shear stresses uv and uw (both have a dash above them). Can i still use your file ?

In general i have to evaluate using Ansys an experiment which took place 30 years ago. And i only have the resulted contours. Nothing more. But i believe the people who conducted the experiment defined the uu Reynolds stress as u'*u' with no density.

[LuckyTran]

Well then you shouldn't call those Reynolds stresses because they are not stresses.

You can use the script but you need to remove the density. Anyway, 99% of the work has already been done for you, good luck. The problem with the script is it's written in language that you must decipher. It's much faster to start from scratch.

If you know the theory behind Boussinesq and you can write any custom field function, you can skip my script and start over from scratch. You will be done faster this way. You will end up with a very similar script.

You should just go ahead and do it instead of wondering what-if. If you had done this from the start, the problem would have been solved days ago already.